Feedbucket MCP Server
Enables AI assistants to manage website feedback from Feedbucket, including listing, filtering, commenting on, and resolving feedback items through natural language interactions.
README
Feedbucket MCP Server
A Model Context Protocol (MCP) server that lets AI assistants (Claude Code, Cursor) manage website feedback from Feedbucket.
Beta Software: This project is in beta. Use at your own risk. Please report issues on GitHub.
Prerequisites
Before installing this MCP server, make sure you have:
- A Feedbucket account
- Feedbucket installed on your website
- Protected mode enabled (recommended):
- Go to Widget Settings
- Set "Trigger Feedbucket using a query string" to Yes
- Save settings and copy your secret key
Quick Start
1. Install from npm
npm install -g feedbucket-mcp
2. Run the setup wizard
# For protected projects:
feedbucket-setup https://your-website.com YOUR_FEEDBUCKET_SECRET
# For public projects:
feedbucket-setup https://your-website.com
The setup script automatically:
- Extracts the project ID from your website HTML
- Fetches the private key from Feedbucket API
- Configures Claude Code and Cursor
3. Restart your IDE
Restart Claude Code or Cursor to load the new MCP server.
4. Start using it
Ask your AI assistant:
- "Show me unresolved feedback from Feedbucket"
- "What's the feedback summary for this project?"
- "Add a comment to feedback #12345 saying we've fixed this"
Alternative: Clone from GitHub
If you prefer to clone the repository:
git clone https://github.com/swiftcomza/feedbucket-mcp.git
cd feedbucket-mcp
npm install && npm run build
npm run setup -- https://your-website.com YOUR_FEEDBUCKET_SECRET
Available Tools
| Tool | Description |
|---|---|
feedback_list |
List and filter feedback with smart summarization |
feedback_get |
Get full details of a specific feedback item |
feedback_stats |
Quick project overview (total, resolved, by type, recent activity) |
feedback_comment |
Add comments to feedback items |
feedback_resolve |
Mark feedback as resolved |
api_status |
Check API connection status |
feedback_list
List feedback with powerful filtering:
Parameters:
- resolved (boolean): Filter by resolution status
- limit (number): Items to retrieve (default: 10, max: 50)
- offset (number): Skip items for pagination
- summary (boolean): Return condensed format (default: true)
- page_filter (string): Filter by page URL (partial match)
- reporter_filter (string): Filter by reporter name
- feedback_type (string): 'screenshot', 'video', or 'text'
- created_after (string): ISO date filter
feedback_get
Get complete details for a single feedback item:
Parameters:
- feedback_id (number, required): The feedback ID
Returns full text, all comments, attachments, browser/device info, and console logs.
feedback_stats
Get a quick project health overview:
Returns:
- Total feedback count
- Resolved vs unresolved breakdown
- Feedback by type (screenshot/video/text)
- Recent activity (7 days, 30 days)
- Top 5 pages with most feedback
feedback_comment
Add a comment to a feedback item:
Parameters:
- feedback_id (number, required): The feedback ID
- comment (string, required): Comment text
- reporter_name (string): Defaults to "Claude AI Assistant"
- reporter_email (string): Defaults to "claude@anthropic.com"
- resolve (boolean): Also resolve the feedback (default: false)
feedback_resolve
Mark a feedback item as resolved:
Parameters:
- feedback_id (number, required): The feedback ID
Setup Options
Automatic Setup (recommended)
# Protected project - provide URL and secret
npm run setup -- https://your-website.com YOUR_SECRET
# Public project - just the URL
npm run setup -- https://your-website.com
Manual Project ID (if auto-detection fails)
# With project ID only (public)
npm run setup -- --project-id tMbUCuQ4FeFJVViMEPlb
# With project ID and API key (protected)
npm run setup -- --project-id tMbUCuQ4FeFJVViMEPlb --api-key YOUR_SECRET
Configure only one IDE
# Claude Code only
npm run setup -- https://your-website.com YOUR_SECRET --claude
# Cursor only
npm run setup -- https://your-website.com YOUR_SECRET --cursor
Extract credentials without configuring
npm run setup -- https://your-website.com YOUR_SECRET --extract
Full help
npm run setup -- --help
Manual Configuration
If you prefer to configure manually:
Claude Code
Add to .mcp.json in your project root:
{
"mcpServers": {
"feedbucket": {
"command": "node",
"args": ["/path/to/feedbucket-mcp/dist/index.js"],
"env": {
"FEEDBUCKET_PROJECT_ID": "your-project-id",
"FEEDBUCKET_PRIVATE_KEY": "your-private-key",
"FEEDBUCKET_API_KEY": "your-api-key"
}
}
}
}
Or use the CLI:
claude mcp add feedbucket \
-e FEEDBUCKET_PROJECT_ID="your-project-id" \
-e FEEDBUCKET_PRIVATE_KEY="your-private-key" \
-e FEEDBUCKET_API_KEY="your-api-key" \
-- node /path/to/feedbucket-mcp/dist/index.js
Cursor
Add to .cursor/mcp.json in your project root:
{
"mcpServers": {
"feedbucket": {
"command": "node",
"args": ["/path/to/feedbucket-mcp/dist/index.js"],
"env": {
"FEEDBUCKET_PROJECT_ID": "your-project-id",
"FEEDBUCKET_PRIVATE_KEY": "your-private-key",
"FEEDBUCKET_API_KEY": "your-api-key"
}
}
}
}
Environment Variables
| Variable | Required | Description |
|---|---|---|
FEEDBUCKET_PROJECT_ID |
Yes | Project ID from the data-feedbucket attribute |
FEEDBUCKET_PRIVATE_KEY |
Yes | Private key for commenting/resolving (auto-fetched by setup) |
FEEDBUCKET_API_KEY |
No* | API key for protected projects (feedbucketKey from URL) |
*Required if your project uses query string protection.
Development
# Install dependencies
npm install
# Development mode (with hot reload)
npm run dev
# Type checking
npm run typecheck
# Linting
npm run lint
# Build for production
npm run build
Architecture
This MCP server was reverse-engineered from the Feedbucket browser widget to provide:
- Smart Summarization: Optimizes large datasets for AI consumption
- Client-side Filtering: Reduces API calls by filtering locally
- Full TypeScript: Strict mode with comprehensive types
- Production Ready: Proper error handling and validation
API Endpoints Used
| Method | Endpoint | Purpose |
|---|---|---|
| GET | /projects/{id} |
Fetch project with all feedback |
| POST | /feedback/{id}/comments |
Add comment to feedback |
| PUT | /feedback/{id}/resolve |
Mark feedback as resolved |
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.